目的系统阐明全氟辛酸（PFOA）暴露致结直肠癌（CRC）的分子毒理机制，筛选核心靶点并验证其表达特征、免疫相关性及临床价值，构建PFOA诱导CRC的不良结局通路（AOP）假说框架。方法整合网络毒理学、多机器学习算法（LASSO回归、SVM-RFE、随机森林）及多维生物信息学分析。从公共数据库收集PFOA与CRC相关靶点，筛选交叉靶点并进行功能富集分析；通过GEO和TCGA（UALCAN平台）数据集验证核心靶点的差异表达、免疫浸润及预后意义；采用分子对接及100 ns分子动力学模拟评估PFOA与核心靶点的结合稳定性；基于关键事件关联构建AOP框架。结果共鉴定738个PFOA-CRC交叉靶点，显著富集于凋亡通路、脂质与动脉粥样硬化通路及CRC通路。机器学习筛选出6个核心靶点[ATP结合盒亚家族G成员2（ABCG2）、紧密连接蛋白1（CLDN1）、ETS变异转录因子4（ETV4）、溶质载体家族7成员5（SLC7A5）、MMP7、CA9）]，这些基因在不同队列CRC样本中表达持续失调（P＜0.05），且与免疫细胞浸润密切相关： ABCG2增强抗肿瘤免疫应答，CLDN1、ETV4、SLC7A5、MMP7和CA9协同促进免疫抑制微环境形成。临床验证显示核心靶点表达与CRC患者预后仅呈关联趋势，未达到统计学显著性（P＞0.05）。分子对接证实PFOA与所有核心靶点稳定结合，其中与CLDN1结合力最强（-29.34 kJ·mol^-1），分子动力学模拟进一步验证了复合物的稳定性。结论 PFOA通过调控多核心靶点、重塑肿瘤免疫微环境促进CRC进展。本研究构建的整合计算框架及AOP假说，为环境污染物健康风险评估及CRC靶向干预提供了理论基础和实验依据。

Objective To systematically clarify the molecular toxicological mechanisms of perfluorooctanoic acid (PFOA) exposureinduced colorectal cancer (CRC), identify core targets, verify their expression characteristics, immune relevance, and clinical value, and construct a hypothetical AOP framework for PFOA-induced CRC. Methods Network toxicology, multiple machine learning algorithms (LASSO regression, SVM-RFE, and Random Forest), and multidimensional bioinformatics analyses were integrated. PFOA-related and CRC-related targets were collected from public databases, and overlapping targets were screened and subjected to functional enrichment analysis. The differential expression, immune infiltration, and prognostic significance of core targets were validated using GEO and TCGA datasets (via the UALCAN platform). Molecular docking and 100-ns molecular dynamics (MD) simulations were performed to evaluate the binding stability between PFOA and core targets. An AOP framework was constructed based on the association of key events. Results A total of 738 overlapping PFOA-CRC targets were identified, which were significantly enriched in pathways such as “apoptosis”, “lipid and atherosclerosis”, and “colorectal cancer”. Six core targets (ABCG2, CLDN1, ETV4, SLC7A5, MMP7, and CA9) were screened by machine learning. These genes showed consistent dysregulation in CRC samples from different cohorts (P < 0.05) and were closely associated with immune cell infiltration: ABCG2 enhanced anti-tumor immune responses, while CLDN1, ETV4, SLC7A5, MMP7, and CA9 synergistically promoted the formation of an immunosuppressive microenvironment. Clinical trials demonstrated that core target expression showed only an association trend with prognosis in CRC patients, failing to reach statistical significance (P > 0.05). Molecular docking confirmed stable binding between PFOA and all core targets, with the strongest binding affinity observed for CLDN1 (-29.34 kJ·mol^-1), which was further validated by molecular dynamics simulations confirming the stability of the complex. Conclusion PFOA promotes CRC progression by regulating multiple core targets and remodeling the tumor immune microenvironment. The integrated computational framework and AOP hypothesis constructed in this study provide a theoretical basis and experimental evidence for the health risk assessment of environmental pollutants and targeted interventions for CRC.

R285.5

第四届国医大师传承工作室（中国中医药办人教函2022 245号1123-04）；国家中医药传承创新中心的建设（ 202203）；山西省中医院管理局项目（ zyytd2024001）；山西省中医院管理局项目（ zyyyjs2024001）；山西省卫健委山西省医学科技创新团队项目（ 2020TD04）；山西省卫健委山西省医学重大科技攻关专项（2022XM10）